lena lavinas

21ST CENTURY WELFARE

L
atin america has long served as a proving ground for eco-
nomic and political experiments that later acquire a global
reach: the shock therapy of neoliberalism was followed by
structural adjustment programmes that were visited on debt-
stricken states across the continent in the 1980s, before being rolled out
in Africa and elsewhere.1 Since the late 1990s, the region has also served
as the laboratory for what the Economist has called ‘the world’s favourite
new anti-poverty device’: conditional cash transfer programmes (ccts)
which, as their name suggests, supply monetary benefits as long as recip-
ients can demonstrate that they have met certain conditions. In 1997,
only three Latin American countries had launched such programmes;
a decade later, the World Bank reported that ‘virtually every country’ in
the region had one, and others outside it were adopting them ‘at a pro-
digious rate’. By 2008, 30 countries had them, from India, Turkey and
Nigeria to Cambodia, the Philippines and Burkina Faso; even New York
City had put one in place.2

The reasons for this proliferation appear simple. ccts hold out the
prospect of killing several developmental birds with one stone: by tying
receipt of benefits to children’s attendance at school or to family vis-
its to health centres, they aim to reduce extreme income poverty while
also addressing other disadvantages suffered by the poor—rectifying
what development-speak calls ‘underinvestment in human capital’. In
many cases they also claim to advance an agenda of ‘female empower-
ment’, either by requiring women to be the recipients of the cash or by
making girls’ education a condition of disbursement. Further, by ‘target-
ing’ recipients and imposing conditions, ccts offer a way to attenuate
extreme poverty without imposing the kind of fiscal burdens that univer-
sal welfare provision would involve; they are an ad hoc benefit, subject
to significant budget constraints. The Economist concluded approvingly
in 2010 that ‘the programmes have spread because they work. They cut

new left review 84 nov dec 2013 5

 6 nlr 84

poverty. They improve income distribution. And they do so cheaply.’3

Little wonder, then, that governments across the developing world,
policy experts and multilateral financial institutions—the World Bank
foremost among them—have increasingly turned to such programmes
as their weapon of choice in the ‘war on poverty’.

The rise of ccts has unfolded in the midst of a broader shift in the
nature of social protection, affecting global South and wealthy North
alike. In many rich industrialized states, governments of both centre-
right and centre-left have proclaimed that they can no longer afford
universal welfare systems of the kind created during the twentieth
century. Over the last three decades, many have moved to downsize or
dismantle them, shifting from comprehensive coverage towards more
individualized models—‘targeted’ or ‘means-tested’—and from decom-
modified provision of goods and services to a greater emphasis on cash
benefits. The differences are by no means trivial, underpinned by an
ideological sea change with far-reaching effects. Whereas one function
of the post-war welfare state had been to remove core provision of health,
education, housing and social insurance from the buffetings of the mar-
ket, the role of the new-model ‘enabling state’ is to facilitate the play
of market forces—providing ‘public support for private responsibility’.4
Rather than recognizing needs, it concedes ‘entitlements’, and instead
of ensuring equal access to public goods, it offers rewards in exchange
for the fulfilment of obligations—the quintessential coinage in this
sense being ‘workfare’.

In the West, one of the key mechanisms for promoting individual

responsibility has been financialization: the expansion of credit markets
enables citizens better to ‘manage risk’, with personal and household
debt serving in theory both to liberate citizens from dependency on a
retreating state and to discipline the feckless. These same doctrines of

1
An early draft of this article appeared as ‘Latin America: Anti-Poverty Schemes
Instead of Social Protection’, desiguALdades Working Paper no. 51, 2013. I thank
Verónica Schild, Robert Boyer, Sérgio Costa, Barbara Fritz and other fellows for
their critical comments during my stay at desiguALdades in the autumn of 2012; I
am grateful to Tatiana Ferro, Francisca Talledo, Flora Thomson-DeVeaux and Paul
Talcott for their valuable assistance.
2
Economist, 29 July 2010; Ariel Fiszbein, Norbert Schady et al, Conditional Cash
Transfers: Reducing Present and Future Poverty, Washington, dc 2009.
3
Economist, 29 July 2010.
4
Neil Gilbert, The Transformation of the Welfare State, Oxford 2002, p. 4.
 lavinas: Welfare 7

individual responsibility and risk management have also been advanced

across much of the global South, most prominently by international
financial institutions, development agencies and ngos. Here the
agenda has been driven not so much by a desire to dismantle universal-
ist mechanisms—countries in the developing world generally lacked the
comprehensive social insurance schemes that were a feature of the Cold
War in the West—as by a twofold emphasis on economic growth and
‘human capital accumulation’. The generally low educational levels and
vulnerable health of the poor are seen as an obstacle to prosperity, not
least because they prevent them from participating fully in the market.
As one imf functionary emphatically asserted at a seminar co-organized
by the Friedrich-Ebert-Stiftung and ilo, ‘there is no vibrant economy if
there are no consumers.’5 In this agenda, the battle against poverty and
the advance of finance-led capitalism have fused.

In the 1980s and 1990s, the tools of choice for integrating the deserving
poor into the market were microcredit schemes, such as Grameen Bank
in Bangladesh or BancoSol in Bolivia. Despite many enthusiastic claims
made for them, the impact of such schemes on poverty rates was mod-
est, to say the least.6 Since the turn of the century, thanks to their record
of apparent success in Latin America, it is ccts that have moved to the
fore. Such programmes are not merely a technical device for combating
poverty. By targeting recipients on condition that they demonstrate ‘co-
responsibility’ for their own welfare, the schemes reinforce the trend
away from universal provision and towards a limited, ‘residual’ model of
social protection. At the same time, by providing select groups of the poor
with cash or new modalities of bank credit rather than decommodified
public goods or services, they are also a powerful instrument for draw-
ing broad strata of the population into the embrace of financial markets.
In that sense, the global spread of ccts is part of a wider reshaping of
welfare regimes in the developing world and beyond.

But just how effective have ccts been in reducing poverty, and what
have been their wider consequences for social provision in countries
that have adopted them? The experience of Latin America, where the

5
Quotation from Elliott Harris, fes–ilo Seminar on the Social Protection Floor,
Berlin, November 2012.
6
For a robust comparative study of microcredit schemes carried out during the
1980s and 90s, see David Hulme and Paul Mosley, eds, Finance Against Poverty,
vols I and II, London 1996.
 8 nlr 84

policy was developed and road-tested on populations from Mexico City

to Santiago, from the Brazilian sertão to the Peruvian altiplano, offers the
broadest range of case studies to date. In what follows, I trace the emer-
gence and take-up of ccts across the region, and examine the evidence
on their outcomes.

The decisive impetus for the design and implementation of new safety
nets came from the severe fiscal and economic crises of the 1980s. The
debt spirals that resulted from the hiking of us interest rates in 1979
brought high inflation, unemployment and sharp declines in real wages
across Latin America, as growth stalled for what became known as the
‘lost decade’. The remedies applied—imf-decreed structural adjustment
plans, which involved sweeping cuts in social spending and elimination
of subsidies—aggravated the situation, deepening levels of destitu-
tion and forcing millions into the informal economy. Over the course
of the 1980s, Latin America witnessed a significant increase in poverty
and ‘indigence’ (extreme poverty) rates: according to figures from the
Economic Commission for Latin America and the Caribbean (eclac),
the overall poverty rate for the region went from 41 per cent in 1980
to 48 per cent in 1990, with indigence rates rising from 19 to 23 per
cent. The number officially classed as poor reached 204 million people
in 1990, as against 136 million ten years earlier.

There was clearly an urgent need for some sort of cushion against the
consequences of liberalization. The existing pay-as-you-go social pro-
tection systems, largely the privilege of formal-sector employees, were
unable to cope with the effects of structural adjustment, and those out-
side them fared still worse. But the solutions sought for this situation
during the 1990s involved not a reversal, but an extension of the neo-
liberal paradigm, to which many governments had converted radically
and abruptly, pushing through extensive and rapid privatization pro-
grammes. Two strategies were pursued initially. On the one hand, the
public pension systems were to be fully or partly privatized, to reduce
the fiscal burden imposed by demographic shifts (population ageing)
coupled with low growth and high rates of informality among the work-
ing population. Several Latin American countries adopted pension
reforms which, following the example set by Chile in the early 80s,
entailed an expansion of the private sector’s role: Mexico and Peru in
1992, Argentina and Colombia in 1993, Uruguay in 1995, Bolivia in
1996. A central aim was to foster the development of capital markets
 lavinas: Welfare 9

in Latin America, considered relatively feeble at this point. On the other

hand, at the same time as it withdrew from the social responsibilities
of pension provision, the ‘enabling’ state would play a greater role in
ensuring the smooth operation of markets. Reducing poverty and indi-
gence was a key goal of this strategy, since high levels of destitution
represented a threat to liberalization. Otherwise, who would pay for the
new services to be delivered by the private sector—pensions, health,
electricity, water, communications?

These twin strategies—privatization, marketization—were pur-

sued in parallel during the 1990s, without being integrated into a
single, coherent model. Moreover, the results of this wave of social-
insurance privatization fell far below expectations: as the World Bank
itself acknowledged a decade later, the reforms failed to improve cover-
age rates.7 Partly due to the dismantling of earlier, fragmented public
pension regimes, poverty grew in the 1990s in several countries: Bolivia,
Ecuador, Peru and Venezuela all experienced rises in the poverty rate.
The continuing vulnerability of large sectors of the population, and the
deepening income deficits wrought by the crises of the 1980s and ensu-
ing structural reforms, prompted the development of a different kind
of safety net.

A new model

Conditional cash transfers are often described as originating in Latin

America—an ‘endogenous innovation’, in the proud words of two Inter-
American Development Bank economists.8 The story of their emergence
and spread across the region usually begins with the programmes imple-
mented in Brazil and Mexico in the late 1990s. However, their intellectual
antecedents can be found further north. Conceptually, we might see ccts
as a confluence of two sets of ideas: the idea of ‘human capital’ on the one
hand, and of ‘targeting’ welfare spending on the other. If Chicago School
economics was the founding matrix of the former, the latter took shape
under the influence of behaviourist economics and ‘decision theory’,
as embodied by rand Corporation reports from the late 60s. Earlier

7
Indermit Gill et al, Keeping the Promise of Social Security in Latin America,
Washington, dc 2004, p. xviii.
8
Marco Stampini and Leopoldo Tornarolli, ‘The Growth of Conditional Cash
Transfers in Latin America and the Caribbean: Did They Go Too Far?’, idb Policy
Brief, Nov 2012.
 10 nlr 84

in the decade, Robert McNamara had famously commissioned rand

analysts to write reports for the Pentagon applying economic thinking
to various aspects of military strategy. Daniel Ellsberg’s contributions are
the best known of these, but they also included a 1966 paper on the
‘Economic Theory of Alliances’ co-written by Mancur Olson and Richard
Zeckhauser. Olson had just codified the ‘free rider’ problem in his Logic
of Collective Action (1965), and here he and Zeckhauser, then a PhD stu-
dent at Harvard, applied similar reasoning to the uneven distribution of
defence spending among nato countries—small states ‘free-riding’ on
the us. Zeckhauser soon moved on to the problem of welfare, writing a
rand report in 1968 which asked: ‘How should assistance programmes
to the poor be structured so as to maximize the utility function of the
representative citizen?’ The answer was ‘targeting’, for example by
encouraging the poor to work through tax incentives—something
Zeckhauser recommended to the Nixon Administration in 1970, influ-
enced by Milton Friedman’s ideas on a ‘negative income tax’.9 Positive
incentives were only one form of targeting, however: Zeckhauser sub-
sequently suggested that allocation of transfers could also be improved
by imposing ‘restrictions on recipients’. In order to qualify, recipients
would have to meet certain ‘deadweight costs’, heart-warmingly referred
to as ‘ordeals’: ‘demeaning qualification tests and tedious administrative
procedures’, for example, or a work requirement that meant accepting
precarious, ‘menial’ jobs with low wages.10

ccts are founded on this principle of targeting, but with a philanthropic

twist: the ‘costs’ imposed on recipients—education, health-centre
visits—are actually beneficial to them in the long run. This second com-
ponent of ccts owes much to the work of Chicago School economists
T. W. Schultz and Gary Becker on ‘human capital’, seen as a crucial
‘input’ that explained much of a country’s developmental success. The

9
Mancur Olson and Richard Zeckhauser, ‘An Economic Theory of Alliances’, rand
Corporation memorandum rm–4297–isa, October 1966; Zeckhauser, ‘Optimal
Mechanisms for Income Transfers’, rand Corporation paper P–3878, 1968;
Zeckhauser and Peter Schuck, ‘An Alternative to the Nixon Income Maintenance
Plan’, Public Interest, Spring 1970, pp. 120–30 (the latter paper warmly thanks
Milton Friedman). Zeckhauser later noted with some satisfaction that the idea of
Earned Income Tax Credits was eventually taken up by the Ford Administration:
Schuck and Zeckhauser, Targeting in Social Programs: Avoiding Bad Bets, Removing
Bad Apples, Washington, dc 2006, p. 160, n. 13.
10
Albert Nichols and Richard Zeckhauser, ‘Targeting Transfers through Restrictions
on Recipients’, American Economic Review, vol. 72, no. 2, 1982, pp. 372–7.
 lavinas: Welfare 11

logic of this, of course, was to downplay the role played by structural

factors in keeping underdeveloped countries poor, instead focusing on
the internal failings of the countries themselves—thus offering a coun-
ter to dependency theory. According to one account, ‘Schultz’s ideas on
human capital are essential to the understanding of the history of the
Chicago School expedition to Latin America’, since they had a ‘lasting
impact on the perspective of the us government aid programmes and on
the work developed by American foundations in the area’.11 In particular,
human capital was the ‘banner’ under which Chicago School ideas were
implanted in Chilean universities in the 1960s, strongly influencing the
economists who would devise Pinochet’s drastic liberalization agenda.
Among them was Miguel Kast, who trained in Chicago from 1971–73,
before returning to work at Odeplan, the Chilean state planning agency;
there he carried out extensive work on poverty, producing a national
map of extreme poverty in 1975. This would provide the analytical foun-
dations for the ‘focused’ anti-poverty measures he implemented after
becoming Minister of Labour and Social Security in 1980.12

In this respect as in others, Pinochet’s Chile was the precursor: not only
was it the first Latin American country to fully privatize the adminis-
tration of its pension funds in 1980, it also pioneered the conditional
safety net, establishing the Subsidio Único Familiar in August 1981.
Combining the ideas of human capital with the principles of targeting,
it provided a stipend equivalent to $6 per month to indigent mothers
with school-age children, conditional on school attendance, to pregnant
women and to women with care responsibilities for disabled people. It
was a small-scale programme: less than a thousand beneficiaries were
reached for a total cost of 0.09 per cent of gdp. In the following dec-
ade Argentina also experimented with a cash transfer programme,
introducing the Programa Nacional de Becas Estudiantiles in 1997,
focused on adolescents from poor backgrounds and again conditional
on school attendance. But it was in Brazil and Mexico that income-
support schemes were first extended on a large scale, and the copious

11
Juan Gabriel Valdés, Pinochet’s Economists: The Chicago School in Chile, Cambridge
1995. Valdés studied at the Universidad Católica in Santiago, a key bridgehead for
the ‘Chicago Boys’, in the late 60s; in exile during the dictatorship, he has subse-
quently served Concertación governments as foreign minister (1999–2000) and
diplomat, overseeing the minustah occupation of Haiti from 2004–06.
12
An early document of his thinking on social policy can be found in ‘Política
económica y desarrollo social en Chile’ (1976), in Hernán Burdiles, ed., El pensami-
ento de Miguel Kast en perspectiva, Santiago 2006, pp. 151–60.
 12 nlr 84

documentation and data derived from study of them helped propel the
adoption of ccts elsewhere. Although the cash transfer programmes
implemented in these two countries converged in their declared aims—
short-term poverty relief, coupled with efforts to break intergenerational
cycles of poverty via ‘human capital accumulation’—their origins and
trajectories were distinct.

In Brazil, cash transfer programmes emerged primarily at the municipal

and state level, only later being adopted nationwide. With the political
loosening of the second half of the 1980s, centre-left governments were
elected to run a number of local authorities, mainly in densely populated
metropolitan areas. Thanks in large part to the decentralization prin-
ciples enshrined in the 1988 constitution, such municipalities became
hotbeds of institutional and policy innovation, putting into action ideas
that had been debated by activists, scholars and policymakers during
the preceding years of widespread political mobilization. Initiatives such
as participatory budgeting, made famous by Porto Alegre, and the anti-
hunger campaign of the Programas de Segurança Alimentar originated
in this ferment, as did the country’s first large municipal minimum-
income programme, established in Brasilia in 1995. The Bolsa Escola
provided a monetary stipend to poor families with children between the
ages of 7 and 14, tied to school attendance. In addition to attenuating
poverty, the scheme aimed to reduce drop-out rates and thereby help to
eliminate child labour.

The Bolsa would be held up as a model to be emulated in the rest of the

country, for three main reasons. Firstly, the poverty threshold used to
identify potential beneficiaries was set at per capita family income of half
the minimum wage. Secondly, the benefit was a flat rate that amounted
to a minimum wage, a significant sum by local standards—especially
given that there had never been a policy in Brazil that had specifically
addressed poverty. Finally, the take-up rate was surprisingly high: about
80 per cent of the target population was covered. In view of this, and the
low operational cost, local cash transfer schemes conditional on school
attendance spread rapidly across Brazil; by the end of the 1990s, around
100 municipalities had implemented one. Faced with this evidence,
Cardoso decided to extend the programme nationwide. However, the
attempt at scaling up would fail: no more than 1 million poor families,
barely 10 per cent of the target population, had been enrolled by the end
of his term in 2002. The programme was also redesigned by the federal
 lavinas: Welfare 13

government, losing much of its effectiveness: the poverty line was set at
an even lower level, thereby excluding the bulk of potential beneficiaries,
and the payment was reduced and tailored to distinct age groups, provid-
ing much less assistance to poor families.

In Mexico, by contrast, the first cct scheme was a top-down initia-

tive, designed and implemented by the federal government. Created in
1997, Progresa—the Programa de Educación, Salud y Alimentación—
was a national programme combining education, food and health
benefits, aimed principally at poor rural families. Its main architect was
Santiago Levy, deputy finance minister in the Zedillo government, who
proposed the scheme as a monetary replacement for existing subsidies
on milk, tortillas and other staples. Instead, beneficiaries would receive
a basic monthly food grant and further cash conditional on children’s
school attendance. One of Progresa’s innovations was to establish a
higher stipend for girls, whose drop-out rates were worse than those
of boys, since they were often required to help their mothers with
domestic work. A second novelty was that larger stipends were paid for
children in higher school grades, as an incentive to increase middle-
school enrolment rates. Progresa also differed from previous ccts in its
attention to healthcare: in addition to school attendance, benefits were
conditional on regular family visits to clinics for preventive purposes
(prenatal care and child nutrition). Yet despite this apparent concern
with the population’s long-term wellbeing, health-related activities
amounted to only 8 per cent of the Progresa budget in 1999. Had the
Mexican government been committed to a comprehensive, integrated
approach to poverty reduction, the allocation of such a low share of the
budget to general healthcare—especially in the absence of existing pub-
lic provision—could have been seen as an oversight. But the disparity
was not accidental, as the further development of such programmes in
Mexico and elsewhere would indicate.

Take-up

The spread of ccts across Latin America after 2000 was contingent
on three major factors. Politically, the election of a wave of progressive
governments was crucial: starting with Chávez in 1998, through Lula
in 2002, Morales in 2005 and Correa the following year, among oth-
ers, left or centre-left forces arrived in power who were committed to
redressing some of the worst consequences of the preceding decade’s
 14 nlr 84

frenzy of liberalization. The ‘pink tide’ pushed social concerns up the

agenda across the region, making governments of varied political colour­
ations more likely to support anti-poverty initiatives. Second, after the
crashes of the 1990s and early 2000s—the Tequila Crisis of 1994, the
aftershocks of the Asian Crisis, culminating in the Argentine default in
2002—the continent began to experience a period of renewed if uneven
growth. The ongoing housing and credit bubble in the us and other
major Western states, and the expansion of manufacturing in China,
brought a rise in commodity prices, boosting Latin America’s export rev-
enues; after 2008, financial markets across the region received floods
of hot money seeking higher returns in ‘emerging markets’. This gave
governments a margin of fiscal manoeuvre they had previously lacked.

A third critical factor was institutional: after initial scepticism, the World
Bank and other development agencies became eager to promote ccts.
Although the World Bank and imf had played a leading role in pushing
through the privatization of social insurance in Latin America, until the
mid-1990s both bodies consistently opposed any initiatives providing
cash to the needy in developing countries, on the grounds that the poor
are ‘unable to make efficient choices’; furthermore, they were convinced
that governments there lacked the fiscal capacity to guarantee such
safety nets. However, around the turn of the new century, World Bank
economists began to advance a ‘social risk management’ strategy for
the developing world that offered a pro-market approach to combating
poverty. This envisaged ‘public interventions to assist individuals, house-
holds and communities better to manage risk, and to provide support
for the critically poor’.13 Among the instruments recommended were
means-tested safety nets, as well as improving the access of the poor to
‘market-based risk management instruments’, such as microinsurance
and microcredit. The role of the state would be sharply circumscribed,
and that of financial markets expanded.

The World Bank acknowledged that reducing wide income gaps

would boost market economies throughout the developing world.
Yet it remained wary of simply handing cash to the poor. ccts were
instrumental in its change of attitude. Vital roles were played here by
the Inter-American Development Bank, which enthusiastically backed

13
Robert Holzmann and Steen Jørgensen, ‘Social Risk Management: A New
Conceptual Framework for Social Protection, and Beyond’, Social Protection
Discussion Paper 0006, World Bank 2000.
 lavinas: Welfare 15

ccts from early on, and today claims involvement in ‘just about every
one of those programmes’ in Latin America;14 and, perhaps more impor-
tantly, by the International Food Policy Research Institute (ifpri). A
Washington-based think-tank originally set up to promote the Green
Revolution—it claims Robert McNamara and Norman Borlaug among
its ‘founding fathers’—ifpri was invited by the Mexican government to
carry out an independent technical evaluation of Progresa.15 Its enthu-
siastic reports from Mexico and subsequently Brazil provided much of
the evidentiary basis on which World Bank economists concluded that
‘results from a first generation of programmes reveal that this innovative
design has been quite successful in addressing many of the criticisms of
social assistance such as poor poverty targeting, disincentive effects and
limited welfare impacts.’ The early experience of ccts apparently served
to ‘debunk claims that targeted programmes in poor countries are inevi-
tably plagued by leakage and high administrative costs’.16

As several more Latin American states took up the idea, the World Bank
too embraced ccts as a new paradigm for combating poverty that was
compatible with its ‘social risk management’ agenda; within a few years,
it would be funding pilot projects across the developing world. The
Bank’s chairman, James Wolfensohn, claimed to have been ‘very excited’
on first encountering Progresa: ‘It was homegrown, based on solid eco-
nomic and social analysis, comprehensive in approach, and sensitive to
the institutional and political realities of the country. Most impressive of
all, it was designed from the start to have a measurable and sustained
impact.’17 Among other influential voices joining the chorus of approval
was Gary Becker, who praised Progresa in 1999 as a ‘highly successful’
example other developing countries should follow.18

14
See ‘Poverty Alleviation’, in the ‘Social Protection and the idb’ section of the
bank’s site, www.iadb.org.
15
It was reportedly paid $2.5 million for its services: Susan Parker and Graciela
Teruel, ‘Randomization and Social Program Evaluation: The Case of Progresa’,
Annals of the American Academy of Political and Social Science, May 2005, p. 210.
16
Laura Rawlings, ‘A New Approach to Social Assistance: Latin America’s
Experience with Conditional Cash Transfer Programs’, World Bank Social Protection
Discussion Paper, August 2004; and Martin Ravallion, ‘Targeted Transfers in
Poor Countries: Revisiting the Trade-Offs and Policy Options’, World Bank Social
Protection Discussion Paper, May 2003.
17
Wolfensohn, Foreword to Santiago Levy, Progress against Poverty: Sustaining
Mexico’s Progresa–Oportunidades Program, Washington, dc 2006, pp. vii–viii.
18
Gary Becker, ‘“Bribe” Third World Parents to Keep Their Kids in School’, Business
Week, 21 November 1999.
 16 nlr 84

The speed with which ccts were adopted in one Latin American coun-
try after another can be seen from the chronology in Table 1 (below):
whereas four countries had one in 1997, within five years the number
had doubled, rising to 17 by 2009. Moreover, countries that already had
such programmes either expanded and reconfigured them or added
others. In 2002, for example, the Lagos government in Santiago estab-
lished Chile Solidario; the same year, the Fox government in Mexico
rebranded Progresa as Oportunidades and extended it to urban areas,
while in 2003, the Lula government drew the Bolsa Escola together
with other anti-poverty measures from the Cardoso era—food stamps,

Table 1. The spread of ccts in Latin America and the Caribbean

Year Country Name of programme

1981 Chile Subsidio Único Familiar

1997 Argentina Programa Nacional de Becas Estudiantiles
Mexico Progresa / Oportunidades
1998 Honduras Programa de Asignación Familiar
2000 Costa Rica Programa Superémonos / Avancemos
2001 Colombia Familias en Acción
Jamaica path
2002 Chile Chile Solidario
2003 Brazil Bolsa Família
Ecuador Bono de Desarrollo Humano
2005 Dominican Republic Solidaridad
El Salvador Red Solidaria / Comunidades Solidarias
Paraguay Tekoporã / Nopytyvo / Propais II
Peru Juntos
2006 Panama Red de Oportunidades
Trinidad & Tobago Targeted Conditional Cash Transfer Programme
2008 Argentina Asignación Universal por Hijo
Guatemala Mi Familia Progresa / Mi Bono Seguro
Uruguay Asignaciones Familiares
2009 Bolivia Bono Juancito Pinto

Source: Barbara Cobo, Políticas Focalizadas de Transferência de Renda: Contextos e Desafios, São Paulo 2012.
 lavinas: Welfare 17

a school grant, a natural-gas subsidy—combining and expanding them

significantly to create the Bolsa Família.

Although the programmes vary from one country to another, they have
a number of common features. Firstly, the target population is defined
by means-testing or other criteria, such as location in an impoverished
area; the government agency responsible for identifying potential recipi-
ents issues a call for applicants, and then selects beneficiaries. Second,
the benefits are paid on a monthly or bimonthly basis, but subject to
conditions that can include school attendance, health-clinic visits, par-
ticipation in community meetings and other activities. The modalities
of payment have changed over time: Progresa began with wire transfers
but shifted in 2003 to a system based on individual accounts at Bansefi,
a state-owned savings bank; Bolsa Família has from the outset oper-
ated through a debit card linked to an account at the state-owned Caixa
Econômica Federal. A third common feature of ccts is that monetary
benefits are generally paid to wives or mothers, seen as better able to
optimize scarce resources. Fourth, benefits tend to vary according to
family size. Fifth, the programmes are monitored, both to prevent ‘leak-
age’ to the undeserving and to enforce compliance from beneficiaries.
Finally, penalties apply in cases of non-compliance, leading recipient
families to be removed from the official register and lose the stipend.

Within this framework there is considerable range, both in terms of

scope and conditionality; Tables 2 and 3 (overleaf) respectively rank
the programmes according to expenditure and coverage. Brazil’s Bolsa
Família is the world’s largest cct programme in terms of reach and
budget: by December 2012, around 45 million people had benefited from
the scheme—some 23 per cent of the Brazilian population—and annual
spending totalled around 21 billion reais ($10 billion), equivalent to 0.5
per cent of gdp. The smallest programme relative to the overall popula-
tion is perhaps Argentina’s Programa Nacional de Becas Estudiantiles,
which reaches less than 1 per cent of the country’s inhabitants; although
in 2009 the government of Cristina Fernández established another
stipend, the Asignación Universal por Hijo, giving 460 pesos (around
$125) a month to the children of the unemployed, conditional on school
attendance and fulfilment of healthcare requirements. The size of the
benefits varies widely, from a maximum of $130 in Brazil to less than
$10 in Chile, Honduras or Jamaica. The cheapest in terms of expenditure
relative to gdp is El Salvador’s Red Solidaria, accounting for 0.02 per
 18 nlr 84

Table 2. Latin American ccts ranked by spending (% of gdp)

Country Annual cost (as % of gdp)

Ecuador 1.2
Brazil 0.5
Dominican Republic 0.5
Mexico 0.5
Uruguay 0.5
Colombia 0.4
Costa Rica 0.4
Jamaica 0.4
Paraguay 0.4
Bolivia 0.3
Guatemala 0.3
Argentina (auh + pnbe) 0.2
Honduras 0.2
Panama 0.2
Trinidad & Tobago 0.2
Chile (cs + suf) 0.1
Peru 0.1
El Salvador 0.02

Source: Economic Commission for Latin America and the Caribbean, Social
Panorama of Latin America 2010, Santiago 2010, p. 140, Figure iii.9.

cent of an already small gdp. Chile Solidario is perhaps the most intru-
sive in its conditions: to receive a benefit starting at $24 a month before
gradually declining to $11, recipients have to sign a contract committing
them to ‘personalized assistance’ with their health, education, employ-
ment, family life, housing situation and income, monitored through
regular meetings with social workers.

Impacts

There are three major pieces of evidence used to support the broader
case for ccts. Firstly, it is claimed that the intensity of extreme poverty
dropped significantly. According to eclac, extreme poverty rates in
 lavinas: Welfare 19

Table 3. Latin American ccts ranked by coverage

Country Coverage (% of population) Coverage (% of poor population)

Ecuador 44 >100
Brazil 26 85
Colombia 25 57
Mexico 25 63
Guatemala 23 40
Dominican Republic 21 46
Latin America average 19 48
Bolivia 18 32
Uruguay 12 85
Jamaica 11 >100
Panama 11 40
Honduras 9 12
Paraguay 9 13
Argentina (auh + pnbe) 8 46
El Salvador 8 17
Peru 8 21
Chile (cs + suf) 7 52
Costa Rica 3 17
Trinidad & Tobago 2 15

Source: eclac, Social Panorama of Latin America 2010, p. 141, Table iii.1.

Latin America did indeed drop from 19 per cent in 2002 to 12 per cent in
2010.19 Second, the increase in social spending targeting the most des-
titute improved some key indicators relating to poverty. A 2009 World
Bank report, for example, asserts that ‘virtually every programme that
has had a credible evaluation has found a positive effect on school enrol-
ment’; ‘ccts generally have increased the use of education and (some)
health services.’20 Third, advocates of the programmes claim that by pro-
viding new entitlements, they instituted a new relationship between the
state and the indigent, allowing the latter to make novel social demands
on the former.

19
eclac, Social Panorama of Latin America, Santiago 2012.
20
Fiszbein et al, Conditional Cash Transfers, pp. 125, 129, 141.
 20 nlr 84

How should these claims, and the effectiveness of ccts more generally,
be assessed? The first thing to consider is their impact on the scale and
composition of social spending. It is true that total social spending has
risen sharply in Latin America. Between 1990–91 and 2008–09, accord-
ing to eclac, average annual per capita expenditure went from $318 to
$819, and the size of social spending as a share of gdp rose by 6.6 per
cent, accounting for 63 per cent of all public expenditure in 2008–09,
as against 45 per cent in 1990–91. The trend certainly looks very posi-
tive. Nevertheless, this growth has been unbalanced: monetary benefits
have registered greater increases than other modalities of public provi-
sion, such as spending on education, healthcare or housing. As Figure 1
(below) makes clear, monetary income transfers—either contributory,
as in pensions, or means-tested benefits—accounted for over half the
overall increase in public social spending, rising as a share of gdp by
3.5 per cent between 1990–91 and 2008–09. By contrast, spending on
health rose by only 1 per cent over twenty years, and on housing by a
mere 0.4 per cent.

Figure 1: Latin American public spending by sector, 1990–2009 (% gdp)

20
1990–1991 1998–1999 2008–2009
18

16
6.6
14

12

10

6 3.5

4 1.8
1.0
2 0.4
0
Total Social Education Health Security and Housing and
Spending Social Welfare Other Areas

Source: eclac social-expenditure database.

 lavinas: Welfare 21

Looking at Latin American countries individually, we see a pattern of

stagnation or even decline in health spending in the first half of the
2000s, followed by an upturn in several countries after 2005, with the
exceptions of Colombia, Peru and Guatemala (Figure 2, below).

But overall, fundamental areas of social provision have lagged behind

the growth of the region’s economies. Unmet demand in these
areas—healthcare, housing and so on—has had to be offset by private
household spending, reinforcing the role of private providers and the
trend towards commodification of basic rights. Moreover there is a
flag­rant contradiction in governments establishing cct programmes
that require medical visits, when they have made little effort to provide
better public healthcare. In this perverse dynamic, the state’s failure
to ensure adequate provision is occluded, and responsibility for poor
health indicators imposed on those who are supposedly in need of
assistance to improve them.

Figure 2. Public spending on health as % of gdp, 2000–2010

7

6
Guatemala
5 Argentina

4
Uruguay
Brazil Mexico
Chile
3

Colombia
2 Bolivia

Peru
1
Venezuela

0
2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010

Note: data does not include extra-budgetary spending—which would raise the Venezuelan
figure significantly, to between 5 and 6 per cent of gdp. Values calculated by dividing by
current prices in the national currency.
Source: author’s calculations based on eclac, Social Panorama of Latin America 2012.
 22 nlr 84

Closer examination of two cct programmes will allow us to assess more

clearly the claims made on their behalf. Peru’s ‘Juntos’—‘Together’—
programme, begun in 2005, targeted poor families living in rural areas
hit by the ongoing civil conflict, in an effort to pacify areas controlled by
guerrilla groups. Foremost among the eligibility criteria was ‘exposure
to violence’, followed by more conventional indicators such as the sever-
ity of poverty and malnutrition.21 All beneficiary households receive a
monthly stipend of approximately $30, regardless of family size; between
2005 and 2011, Juntos reached around 475,000 households—around 6
per cent of the population—including 1 million children, at a minimal
cost: 0.2 per cent of gdp. Yet an evaluation of the programme made in
2010 by two World Bank economists acknowledged that, although the
programme helped to narrow the poverty gap—they weighed its contri-
bution at 5 percentage points—the monetary benefit was insufficient to
raise all recipients’ incomes as far as the poverty line; its long-term effect
on poverty would thus be limited.22 The additional income did help with
nutrition, allowing beneficiaries access to a better diet on a more regular
basis. But the healthcare impact of the programme was again limited,
because of continuing lack of access to public health services: immuni-
zation rates fell far below the targets, with only half the planned number
of children and pregnant women covered after five years. Finally, the
programme had no discernible impact on educational attainment, since
the reported enrolment rates and school-attendance levels were similar
among beneficiaries and non-beneficiaries alike.

Analogous considerations apply to the case of Guatemala’s cct, launched

in 2008 as Mi Familia Progresa (mifapro). By 2011, the programme was
providing a monthly benefit of around $35 to as many as 862,000 fami-
lies, including 1.6 million children under 15 years old—an estimated 35
per cent of the total population. By this stage the programme was costing
0.36 per cent of gdp.23 Like the Peruvian Juntos, mifapro did not report
the expected outcomes: neither school attendance nor family health cov-
erage improved significantly, once again due to shortages on the supply
side. In 2011, when a conservative coalition came to power, mifapro
was renamed Mi Bono Seguro and drastically scaled down, now reach-
ing only 110,000 families—an eightfold reduction. Overall poverty rates

21
Elizaveta Perova and Renos Vakis, ‘Welfare impacts of the “Juntos” Program in
Peru: Evidence from a non-experimental evaluation’, World Bank, March 2009.
22
Perova and Vakis, ‘Welfare impacts of the “Juntos” Program’, pp. 13–14.
23
undp Guatemala, ‘Ejercicio de apreciación sustantiva: Mi Familia Progresa’,
July 2011.
 lavinas: Welfare 23

have increased in Guatemala as of late, rising from 51 per cent in 2006

to 54 per cent in 2011, according to a national household survey carried
out by the government; though the extreme poverty rate dropped from
15 per cent to 13 per cent over the same period.24

The cases of Peru and Guatemala indicate that, of the three main
outcomes sought by ccts—substantial reduction in the intensity of
extreme poverty; rise in social spending as a percentage of gdp; refor-
matting of social demands on the state by the poor—only the first has
been accomplished to any degree. Moreover, the trends in healthcare
spending shown in Figure 2, above, demonstrate that the Peruvian and
Guatemalan governments did nothing to improve public provision.
Indeed, while Peru’s health spending over the decade was stagnant, that
of Guatemala dropped sharply just before the introduction of mifapro
and did not recover thereafter. In other words, in both of these cases, the
state imposed on beneficiary families the burden of finding non-existent
services in order to prove their ‘responsibility’, and their worthiness to
keep receiving the meagre sums they were being offered.

Limitations

Across Latin America, ccts have varied in their eligibility criteria and
conditionalities, aiming at distinct ‘target populations’; the amounts
of the stipends also vary. Generally speaking, however, these schemes
have had only a modest effect on the vast inequalities for which Latin
America is renowned.25 They share a number of significant limitations,
both in practice and in principle. To begin with the question of ‘target-
ing’, the criteria used to identify potential beneficiaries rely on absolute
indigence and poverty lines that have been set at extremely low levels:
the equivalent of an income of $1 and $2 per day, which is lower than the
indigence and poverty thresholds applied by the World Bank ($1.25 and
$2.50 respectively). This tends to hide the real magnitude and severity of
destitution. Second, in most programmes neither the poverty thresholds
used nor the benefits paid are adjusted annually for inflation; the actual
value of the stipends to recipients thus tends to be eroded over time. In
Brazil, for instance, poverty lines and benefits for the Bolsa Família have

24
encovi (Encuesta Nacional de Condiciones de Vida), 2011. Close to three quar-
ters of the poor in Guatemala are indigenous.
25
Kelly Hoffman and Miguel Angel Centeno have dubbed it ‘The Lopsided
Continent’: Annual Review of Sociology, vol. 29, 2003, pp. 363–90.
 24 nlr 84

not been adjusted for inflation since 2009, flying in the face of guide-
lines stipulating that it should rise in line with other benefits, whose
value is indexed annually. Third, none of these programmes displays
a take-up rate of 100 per cent; far from it, for they suffer from hori-
zontal inefficiencies due to inadequate targeting and means-testing by
the government agencies responsible for them. Often, exclusion from
the system or non-registration is a discretionary decision taken at the
local level. Fourth, the monitoring mechanisms which are supposed to
send information about school attendance and medical visits from the
municipal to the federal level, are frequently inefficient; the vast majority
lack computerized systems to process and analyse incoming data. Fifth,
in countries where universal public elementary schooling was already in
place, such as Brazil or Argentina, no correlation between cash transfer
programmes and increased matriculation was observed.26

Moreover, many of these schemes are funded through general taxa-

tion—to which indirect taxes on consumption often contribute heavily;
this means they are very likely to be regressive, since any rise in the
consumption levels of beneficiaries will in turn contribute to the pro-
grammes’ funding. The relative cheapness of the programmes is another
obvious limitation: all but one of the programmes involve spending less
than 0.5 per cent of gdp—the exception is Ecuador’s Bono de Desarrollo
Humano—and most of them are tiny in absolute terms, down to 0.02
per cent of gdp in El Salvador. Their impact on poverty levels was there-
fore inevitably going to be restricted, given the scale of the problem
across the region. Last but by no means least, all of these cct schemes
operate on a residual basis, as a safety net to compensate for market
failures; no Latin American country has transformed them into rights
guaranteeing a minimum income. They provide some compensation
to the needy, yet they remain disconnected from anti-cyclical and per-
manent redistributive policies—a constitutive element of any system of
universal social protection.

The extent to which cct programmes have actually contributed to reduc-

tions in poverty rates across Latin America has prompted a lively debate,
with recent studies indicating that economic growth and job creation
have had a much greater impact. Comparative, cross-country analyses

26
Lena Lavinas, Barbara Cobo and Alinne Veiga, ‘Bolsa Família: impacto das trans-
ferências de renda sobre a autonomia das mulheres e as relações de gênero’, Revista
Latinoamericana de Población, vol. 6, no. 10, 2012, pp. 31–54.
 lavinas: Welfare 25

demonstrate that increased wage earnings account for as much as half

the reduction in poverty across the developing world.27 Similarly, in
Latin America and the Caribbean, according to eclac, ‘in the countries
where poverty lessened, labour income accounted for half or more of the
change in total per capita income’; transfers, both public and private, and
other income contributed ‘to a lesser degree’.28 Among the fundamental
mechanisms that have driven reductions in poverty and labour-market
inequality and boosted consumption in the region, the revalorization
of the real minimum wage would seem to have been crucial: Figure 3
(overleaf) shows a broad recovery from the lows of the 1980s and 90s in
most countries—with the notable exceptions of Mexico, where the trend
is static, and Venezuela, where it is more erratic. Argentina, Bolivia,
Brazil and Ecuador, where the growth in real minimum wages since
2000 has been strongest, are not coincidentally among the countries
which have produced the largest reductions in poverty: according to
eclac data, between 2002 and 2010 the poverty rates in these coun-
tries dropped by 26, 20, 13 and 12 percentage points respectively. Only
Peru, Venezuela and Colombia—countries where booming commodity
prices fuelled significant growth—could boast of comparable reductions
in poverty during the same period, of 23, 21 and 12 percentage points
respectively.29 Conversely in Mexico, whose much-praised cct scheme
has been operating for more than 15 years, poverty fell by only 2 per cent
over the period 1992–2010, according to official estimates.30 As a matter
of fact, between 2008 and 2010, the poverty rate rose from 45 to 46 per
cent, increasing the headcount to 52 million people.

The Brazilian case

The Bolsa Família has been widely touted as a success story. Would an
assessment of its actual impact differ radically from the picture of ccts
elsewhere in Latin America presented above? Initially introduced in

27
Gabriela Inchauste et al, ‘When Job Earnings Are behind Poverty Reduction’,
Economic Premise (World Bank), no. 97, November 2012, found that labour income
accounted for 50 per cent of the reduction in 10 out of 16 countries studied, and for
40 per cent in another 2 countries.
28
eclac, Social Panorama, p. 56.
29
eclac, Social Panorama, pp. 79–80.
30
Figures from Consejo Nacional de Evaluación de la Política de Desarrollo
Social (coneval); see also Luis Rigoberto Gallardo Gómez and David Martínez
Mendizábal, ‘México, la persistente construcción de un estado de malestar’, Revista
de Ciencias Sociales, nos 135–136, 2012, pp. 215–25.
 26 nlr 84

Figure 3. Real minimum wage, average annual index (2000 = 100)

Argentina Bolivia
400 160
140
300 120
100
200 80
60
100 40
20
0 0
1980 1985 1990 1995 2000 2010 1980 1985 1990 1995 2000 2010

Brazil Chile
200 140

120
150
100

80
100
60

40
50
20

0 0
1980 1985 1990 1995 2000 2010 1980 1985 1990 1995 2000 2010

Colombia Ecuador
120 300

100 250

80 200

60 150

40 100

20 50

0 0
1980 1985 1990 1995 2000 2010 1980 1985 1990 1995 2000 2010
 lavinas: Welfare 27

Mexico Peru
350 350

300 300

250 250

200 200

150 150

100 100

50 50

0 0
1980 1985 1990 1995 2000 2010 1980 1985 1990 1995 2000 2010

Uruguay Venezuela
300 250

250
200

200
150
150
100
100

50
50

0 0
1980 1985 1990 1995 2000 2010 1980 1985 1990 1995 2000 2010

160
Latin America (average)

140

120

100

80

60

40

20

0
2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010

Source: eclac database, using official sources. Minimum wages deflated by either the
national or (in the cases of Peru, Mexico and Venezuela) metropolitan consumer price index.
 28 nlr 84

2003, the Bolsa was formally established by law in January 2004, during
Lula’s first term. The programme aims to ensure a minimum monetary
income to poor and indigent families—defined as those with a monthly
per capita family income of 70–140 reais ($35–70) and less than 70
reais ($35), respectively. Rather than providing a single benefit, the pro-
gramme has flexible parameters, adjusting the amount according to the
composition of recipient families. As in most other cases, women are the
nominal recipients of the stipend, effectively acting as the government’s
agents in ensuring compliance with the conditions. In order to receive
the monthly stipend, families are required to make regular visits to health
clinics—aimed especially at pregnant or breastfeeding women, and chil-
dren under five—and to ensure children between the ages of 6 and 17
have a minimum 75 per cent school-attendance record. By December
2012, the Bolsa was being paid to 13.5 million families, a total of around
45 million people—just under a quarter of the Brazilian population.
In geographical terms, the largest concentration of recipients—50 per
cent—is in the Northeast, which has the highest poverty rate in the coun-
try, followed by the Southeast, home to around a quarter of recipients.

Nevertheless, the Bolsa Família shares many of the same limitations

as other cct programmes in Latin America. Again, these range from
technical flaws—inadequacies in the programme’s design or in the tar-
geting mechanism—to the larger question of the Bolsa’s actual effects.
On the first, one important consideration is that, as in other countries,
the Bolsa stipend is not linked to inflation; as a consequence, the recipi-
ents have been getting poorer year by year, since the cumulative inflation
rate from 2009 to 2013 reached almost 25 per cent. The average monthly
stipend amounts to 140 reais or $70 per family. The government also
made a positive step by recognizing that the Bolsa was not reaching all
of those eligible for it. According to estimates released by the Ministry
of Social Development, some 800,000 eligible families—at least 2.5
million people—have not been included in the programme; our own
estimates, based on the National Household Sample Survey conducted
by the ibge, the national statistical body, put the figure much higher,
at 2.2 million families, or 7 million people.31 Two main factors help to
account for this enormous shortfall. Firstly, the targeting mechanism
itself produces inefficiencies, as many potential recipients do not display

31
Lena Lavinas, ‘Pobreza: Métricas e Evolução Recente no Brasil e no Nordeste’, Cadernos
do Des­envolvimento, vol. 5, no. 7, 2010, pp. 126–48.
 lavinas: Welfare 29

the specified characteristics of poverty. For instance, a family in which

one of the members is formally employed and paid a minimum wage,
although its per capita income might fall below the poverty line, is likely
to be ruled out for having some semblance of job security. At the same
time, the imposition of burdens on recipients also serves to reduce take-
up. Secondly, the fact that the Bolsa Família is not a universal right but a
selective welfare benefit, subject to budget constraints, inherently works
to decrease the size of the population it covers.

What of the Bolsa Família’s effectiveness in reducing income poverty?

Here it is important to weigh the impact of the cct relative to wage
earnings and other fiscal transfers, supplied through Brazil’s existing
social security system. By decomposing per capita household income
according to its origin, we can see the contribution to reducing the
poverty rate made by three successive layers of income: (1) wages and
other earnings from paid work (labelled ‘earnings only’); then (2) wage
earnings plus income from pension and other social-insurance benefits
(labelled ‘contributory transfers’); followed by (3) all sources of income,
which includes categories (1) and (2) plus welfare benefits such as the
Bolsa Família and any other earnings. Table 4 (below) displays changes
in the poverty and indigence rates when these three layers of income are
taken into account.

Table 4. Factors in Brazil’s falling poverty and indigence rates

2001 2005 2011

Poverty rate
Earnings only 48 35 26
Earnings + contributory transfers 37 23 15
Earnings + contributory transfers + welfare schemes + other 36 20 11

Indigence rate
Earnings only 28 20 17
Earnings + contributory transfers 17 10 7
Earnings + contributory transfers + welfare schemes + other 16 7 4

Source: Instituto Brasileiro de Geografia e Estatística (ibge).

Estimates based on Bolsa Família poverty lines.
 30 nlr 84

Looking, firstly, at the data for poverty, we can see that in 2001, 48
per cent of the Brazilian population—some 80 million people—were
officially classed as poor if we take only earned income into account.
When we add income received from social-insurance transfers, the pov-
erty rate in 2001 falls to 37 per cent—a decline of 11 percentage points.
This means that, contrary to a widespread prejudice, pensions benefits
in Brazil are not regressive: quite the opposite, given that in 2001 they
lifted the incomes of some 18 million people over the poverty line. The
impact of the third layer of income, however, was much more limited
at this stage, when the system of safety nets remained fragmented
and the Bolsa Família did not yet exist: welfare schemes only reduced
poverty by a further percentage point, benefiting another 2 million peo-
ple. Thus in 2001, 36 per cent of the population lived in poverty, some
60 million people.

By 2011 the picture had changed significantly. The most striking devel-
opment is that, taking only wage earnings into account, the poverty rate
had dropped to 26 per cent—a 46 per cent decrease relative to the 2001
figure—as a direct result of Brazil’s economic growth during this period.
Indeed, according to the data in Table 2, no other source of income
appears to have had as positive an impact on poverty reduction. Thanks
to the new dynamism of the labour market, some 30 million people’s
incomes exceeded the poverty threshold. Moreover, as in 2011, pensions
reduced the poverty rate by a further 11 percentage points, benefiting
21 million more people. The recovery of the minimum wage, whose
value increased by 94 per cent between January 2001 and May 2012, lies
behind both these trends: two-thirds of all public pensions in Brazil cor-
respond to the minimum wage.32 Together, job creation and growth of
the minimum wage brought the poverty rate down to 15 per cent. Finally,
welfare schemes involving cash transfers helped to lower the rate fur-
ther, to 11 per cent, benefiting an additional 7 million people. This is
the lowest share ever recorded since Brazil began keeping household
income data in the mid-20th century. Overall, the poverty rate fell from
36 to 11 per cent in ten years.

A similar analysis can be made of the data on the extreme poverty

rate, which in overall terms went from 16 to 4 per cent over the same

32
The Lula government indexed the minimum wage to changes in the Consumer
Price Index, to allow for inflation relative to the previous year, and also incorporated
the economic growth rate reached two years before.
 lavinas: Welfare 31

period—thus falling by three-quarters. However, here the effects of eco-

nomic growth were not as favourable to those living in extreme poverty
as they had been for those classed simply as ‘poor’. The much lower lev-
els of schooling, and the even more precarious, badly paid jobs held by
the indigent, make them much less likely to benefit from upward trends
in the job market. Thus the indigence rate when only wage earnings are
taken into account fell from 28 to 17 per cent—a 39 per cent drop, com-
pared with 46 for the poverty rate. By contrast, pension benefits were
clearly the major factor in reducing the indigence rate, again thanks to
their indexation to the minimum wage: over the decade, they reduced
the poverty rate by an additional 10 percentage points relative to wage
earnings. Finally, welfare schemes contributed a further 3-percentage-
point reduction, equivalent to 4 million people—a significantly broader
impact than they had had in 2001, thanks to the extension of safety nets,
the Bolsa Família foremost among them.

Nevertheless, the absolute magnitude of poverty remains striking: some

28 million people still fall below the official poverty line. It should also
be recalled that poverty and indigence thresholds in Brazil are set at
extremely low levels; the figures presented above are therefore inevita-
bly underestimates. If Brazil were to implement a poverty line at the
level currently used in the European Union—50 per cent of median
per capita income—the current poverty rate would soar to 40 per cent,
encompassing 70 million people. In 2011, median per capita income in
Brazil amounted to only 466 reais a month, around $240; this in turn
means that two-fifths of the Brazilian population lives with a per capita
monthly income of less than $120. Such figures say a great deal about
the choice of poverty measures in Brazil and other developing countries,
where many commentators have spoken enthusiastically of late about
the emergence of a new middle class.

Alongside the reduction in poverty, there has been a decrease in income

inequality in Brazil over the past decade. Yet it remains staggeringly
high: the country’s Gini index was 0.529 in 2011, compared to 0.593 in
2001. The pattern of income distribution shown in Figure 4 (overleaf)
starkly illustrates the depth and persistence of the disparity: in 2001, the
bottom quintile held barely 2 per cent of all income, compared to more
than 60 per cent for the top quintile; ten years later, the bottom 20 per
cent held just 3 per cent of all income, compared to 57 per cent for the
top quintile. It is worth recalling here the flagrantly regressive nature
 32 nlr 84

Figure 4: Income share in Brazil by quintiles

70

60

50

40

30

20

10

0
2001 2005 2011
Bottom 20% Deciles 3 and 4 Deciles 5 and 6 Deciles 7 and 8 Top 20%

Source: ibge, various years.

of the Brazilian tax system, with its marked incidence of indirect con-
sumption and production taxes, as opposed to direct taxes on income,
inheritance and capital gains. In 2010, the average weight of direct taxes
in oecd countries’ overall tax revenues was 33 per cent, and of indirect
taxes 34 per cent. In Brazil, taxes on income—individual or corporate—
accounted for 19 per cent of tax revenues in 2011, and estate taxes just 4
per cent, whereas indirect taxes accounted for 49 per cent. No product
or service is wholly exempted, which leads to an especially significant
burden for the poorest segments of the population.

As we have seen, it is primarily rising labour earnings that have accounted

for the decline in poverty in Brazil, as was the case elsewhere in Latin
America. Brazil is also no exception to the wider continental tendency
to concentrate social spending on cash transfers rather than expanding
provision of decommodified services, such as public health, education,
sanitation and other basic social goods. Whereas federal social spend-
ing on welfare benefits increased in real terms by 300 per cent between
2001 and 2010, over the same period spending on education doubled
 lavinas: Welfare 33

Figure 5. Households in Brazil with specific facilities and goods (%)

100

90

80 cell phone

70
adequate sanitation
60
clean water for drinking
50
washing machine
40
computer
30

20

10

0
2001 2003 2005 2007 2009 2011

Source: ibge, various years.

and on public health rose by only 60 per cent. The runt of the litter here
is health spending: not only did it grow at a rate below average, but it also
saw its share in federal social spending reduced from 13 per cent in 2001
to 11 per cent in 2010. By that time federal expenditures on education and
welfare benefits amounted to 1 per cent of gdp, whereas sanitation and
housing received only 0.1 and 0.8 per cent of gdp.33 Little wonder, then,
that Brazil ranks so low with respect to living conditions. According to
the ibge data presented in Figure 5 (above), the population’s access to
clean drinking water or adequate sanitation has improved very little over
the last decade. Access to consumer goods such as cell phones, washing
machines and computers, on the other hand, has soared: an amazing 86
per cent of households have at least one cell phone, up from 31 per cent
in 2001, and one in two have a washing machine—though only 2 in 3
households have adequate sanitation. There was no change in the avail-
ability of clean water over the entire decade.

33
ipea, ‘Gasto Social Federal: prioridade macroeconômica no período 1995–2010’,
Nota Técnica no. 9, 2012. It is worth noting, however, that sub-national govern-
ments, such as states and municipalities, also finance education and health through
funds transferred to them from the federal level.
 34 nlr 84

In short, in Brazil as elsewhere in Latin America, social infrastructure

and access to state-provided, decommodified goods and services are
growing at uneven tempos, exacerbating inequalities that are more
difficult to measure than raw labour-income disparities. Patchy state
provision of basic public goods, coupled with rising wage earnings,
have encouraged private spending in education and health. Indeed,
healthcare is a prime example of how a universal right has been dam-
aged by the rationale of finance-led capitalism. The 1988 Constitution
established a right to healthcare, with provision to be ensured by the
state; the Unified Health System (sus) was created in 1990, strongly
influenced by European models such as the British and the French. In
theory, the private sector’s role should be complementary and heavily
regulated by the National Health Agency. In practice, the privatization
of the health system has expanded in the absence of public resources
(though these exist, they have been diverted to other goals). This has
created a vicious cycle of under-financing, steadily worsening since the
sus was founded, which has undermined the system’s universality and
comprehensiveness. In 2009, private spending on health reached 5.3
per cent of Brazilian gdp while public expenditures amounted to only
3.5 per cent. The commodification of health in Brazil seems inexorable,
reflecting the grip of the financial markets.

The dynamic of privatization has been boosted, and the concept of uni-
versality in social provision undermined. A third of the adult Brazilian
population believes that public services should be limited to the des-
titute, and therefore narrowed in scope and quality; although a large
majority—75 per cent—supports some redistribution in favour of the
poor, they do so only if it is tied to conditionalities and controls, with
non-compliance bringing loss of benefits.34 The link between social pro-
vision and selectivity has become strong, as the idea of universal rights
to decommodified public services wanes.

Banks for the bankless

If poverty-reduction has ostensibly been the main motivation for cct

schemes in Latin America, the expansion of the financial sector down
the income hierarchy—what the development literature calls ‘market

34
Lena Lavinas, Barbara Cobo et al., Medindo o Grau de Aversão à Desigualdade da
População Brasileira—um survey nacional, mimeo, November 2012, p. 137.
 lavinas: Welfare 35

inclusion’—has been another important dimension. Indeed, ccts can

be seen as an integral part of a wider drive towards the privatization of
ever larger swathes of economy and society—a process taking in all social
segments, irrespective of income level, which Nancy Fraser has aptly
characterized as ‘commodification all the way down’.35 In Latin America
as elsewhere, financial markets have been central to this endeavour.
As we have seen, earlier programmes extending insurance and credit
to the poor had a modest impact, partly because capital markets in the
countries where they were attempted in the 1980s and early 90s were
weak, during a period of severe structural adjustment and rising pov-
erty. The relative stabilization of Latin America in the early 2000s, and
the effects of the global credit bubble on the region’s economies, altered
the equation. With the advancing financialization of the world economy,
the ‘incomplete’ or ‘missing’ capital markets in low- and middle-income
countries, and their credit sectors in particular, were extended in the
first decade of the 21st century. Increased access to loans at the bottom
of the income pyramid would boost mass consumption, bolstering the
economy from below even as poverty decreased.

Financial markets now assumed a greater role in reshaping the region’s

welfare regimes. The process had begun with the pension reforms of the
1990s, which were in part designed to strengthen Latin America’s stock
markets by putting public funds under private ownership or manage­
ment. But it gathered momentum in the 2000s, as the emphasis on
cash transfers over spending on public goods and services encouraged
individuals and households to seek private alternatives to increasingly
uneven state provision—reinforcing the dynamic towards marketization
at the same time as loans were made available to ever wider strata of the
population. Both income security for the elderly and poverty reduction
were to be achieved through capital markets, which would become the
new providers of welfare, in the form of private insurance on the one
hand, and private credit on the other.

The extension of financial products and services to the poor fits well, of
course, with the World Bank master-concept of ‘social risk management’;
what better way to foster greater responsibility than increased individual
borrowing? However, it requires a level of ‘financial literacy’ that cannot

35
Nancy Fraser, ‘Can society be commodities all the way down? Polanyian reflec-
tions on capitalist crisis’, fmsh Working Papers, no. 18, August 2012.
 36 nlr 84

always be assumed.36 Training programmes and technical advice on the

basic rules of borrowing and how to manage loans, run by ngos and
public institutions, have proliferated in the developing world.37 ccts
have been a key mechanism for the propagation of ‘financial literacy’:
recipients of benefits are often encouraged to attend talks and short-term
courses on the subject. Peru has a pilot programme attached to its Juntos
scheme, run by a ‘financial inclusion’ lobbying group called Proyecto
Capital, which seeks to transmit to families ‘basic knowledge about the
formal financial system and its characteristics, the main products and
services offered, and advantages in terms of security and trust’. The initi-
ative’s website boasts numerous testimonials from contented Peruvians:
‘You feel more alive, because you have your savings, because you can go
to the bank and maybe get a loan in the future’, says one woman; another
confesses that ‘when I kept money in my house, I’d take it out and spend
it whenever I needed something. Now that it’s in the bank, I can’t grab
it as quickly.’38 In Mexico, the bank that disburses Oportunidades pay-
ments, in partnership with two us-based ngos called Freedom from
Hunger and Microfinance Opportunities, has been running workshops
under the rubric ‘Your Money, Your Future’; the objective, it claims, is
to ‘reinforce the behaviours that lead to greater saving, more prudent
spending, justified and manageable levels of indebtedness and a culture
of risk-prevention’.39

Outside Latin America, the gospel of ‘financial inclusion’ is now being

preached in Africa by MasterCard, in initiatives backed by the Bill and
Melinda Gates Foundation, among others. Strikingly, this has involved
a conception of payment technologies as a basic human right; in the
words of Nobel Economics laureate Robert Shiller, the time has come to
‘reframe the wording of universal rights so that they represent the rights
of all people to a fair compromise—to financial arrangements that share

36
Notwithstanding Abhijit Banerjee and Esther Duflo’s observation that ‘the poor
face a huge amount of risk—a friend of ours from the world of high finance reflected
they are like hedge-fund managers’; see Poor Economics: A Radical Rethinking of the
Way to Fight Global Poverty, New York 2011, ch. 6.
37
Lena Lavinas and Camila Ferraz, ‘Inclusão financeira, crédito e desenvolvimento:
que papel uma renda básica pode jogar nesse processo?’, paper presented at bien
13th International Congress, São Paulo, July 2010.
38
See www.proyectocapital.org, ‘Promoción del Ahorro en Familias juntos’
and ‘Testimonials’ sections; the initiative, based in Peru, is funded by the Ford
Foundation and Citibank.
39
See www.bansefi.gob.mx, ‘Educación financiera’ section.
 lavinas: Welfare 37

burdens and benefits effectively. In the future of financial capitalism,

we ought to see better development of our covenants regarding these
“rights”, as financial contracts that are more democratic and nuanced,
with the rights of mankind refined in more basic terms.’40

The integration of increasing numbers of low-income groups into the

financial system has resulted in what has been termed the bancarização—
the ‘bankization’—of the poor, as the gap between the financial sector
and a huge unmet demand for cheap, short-term loans has begun to
be bridged. The Lula years saw a marked expansion of credit in Brazil,
for example, where it rose as a share of gdp from 23 per cent in 2003
to 49 per cent in 2011.41 Much of this derived from the rise in the real
minimum wage noted earlier; but a significant proportion of it was
due to government measures that helped extend various modalities of
credit to poorer segments of the population. Bolsa Família recipients
gained access to special consumption credit lines, such as the Crédito
Fácil, from the Caixa Econômica Federal, which provides loans of up
to $100 with no additional collateral; these are often used to buy dura-
ble goods—refrigerators, tvs, washing machines and so on. There is
also the Construcard, which is designed to support purchases of hous-
ing materials. The big national retailers usually have electronic payment
systems that are integrated with the Caixa Econômica Federal, and loan
requests can be approved almost immediately. The average interest rates
for such loans range from 1.8 per cent to 4 per cent per month; a Crédito
Fácil borrower, for example, would have to pay 4 reais in monthly inter-
est on a 200-real loan, which might seem cheap to someone receiving a
monthly stipend of 130 reais. In parallel with this credit-backed expan-
sion of consumption, there has been a broadening of the supply of other
banking products and services, in particular in the realms of private
insurance. By ‘securing’ needs unmet by the state, the financial system
seems to offer a new kind of social provision.

In the name of the poor

Social policy has long played a marginal role in Latin America: the elites
of the world’s most unequal zone have for centuries ignored those most
in need. In that sense, the ascendancy of ccts over the past decade and

40
Robert Shiller, Finance and the Good Society, Princeton 2012, p. 150.
41
Banco Central do Brasil, Séries Temporais.
 38 nlr 84

a half marks an undoubted shift: most countries in the region today

recognize the need to reduce poverty as a paramount challenge, to be
addressed through large-scale public policies. Even conservative forces
in these countries have been obliged to back schemes which they ini-
tially denounced as clientelist manoeuvres or assumed were doomed to
failure. Hence a broad consensus has been forged around the idea that
ccts are worth implementing—aided by the fact that they are inexpen-
sive, easy to manage and politically rewarding. Yet they remain ad hoc
instruments, unconstrained by legal and institutionalized principles of
rights. The distinction is crucial: instead of being one dimension of a
wider, universal system of social protection, such programmes enforce
a principle of selectivity, targeting the poor as a residual category while
insisting they assume individualized responsibility for their fates—thus
working to diminish social solidarity and cohesion. The schemes are
also designed to extend commodification, on the one hand disbursing
monetary rewards to the poor in exchange for their participation as con-
sumers, while on the other offering governments an alibi for scaling
back provision of public goods. They thus pave the way for a retrench-
ment of welfare rather than its expansion.

A large and diverse body of scholarship has provided ample evidence

that the more universal social-protection systems are, the more redis-
tributive their impact.42 On the basis of such empirical evidence, the
Swedish social scientists Walter Korpi and Joakim Palme famously
identified a ‘paradox of redistribution’, in which the Western welfare
regimes that targeted the poor most heavily actually turned out to
have redistributed much less than expected.43 Evelyn Huber and John
Stephens have recently reaffirmed these findings: Scandinavian coun-
tries stand out as the most effective in reducing poverty and inequality
because they provide large, universal and decommodified services.44 By
contrast, countries whose welfare systems rely principally on means-
tested benefits are much less capable of alleviating poverty and reducing

42
To cite only a few, Gøsta Esping-Andersen, Three Worlds of Welfare Capitalism,
Cambridge 1990; Jonas Pontusson, Inequalities and Prosperity: Social Europe vs
Liberal America, Ithaca, ny 2005; Evelyn Huber and John Stephens, Development
and Crisis of the Welfare State: Parties and Policies in Global Markets, Chicago 2010.
43
Walter Korpi and Joakim Palme, ‘The Paradox of Redistribution and Strategies
of Equality: Welfare State Institutions, Inequality and Poverty in the Western
Countries’, American Sociological Review, vol. 63, no. 5, 1998, pp. 661–87.
44
Evelyn Huber and John Stephens, Democracy and the Left: Social Policy and
Inequality in Latin America, Chicago 2011.
 lavinas: Welfare 39

inequality, as illustrated by the case of the us. According to oecd fig-

ures, member-states with universal social-protection frameworks—the
Nordic countries, France, Belgium, Slovenia—have managed to attain
relatively high levels of equality, with Gini coefficients ranging from
0.27 to 0.32. The strong tendencies towards reciprocity and redistribu-
tion in these societies allowed them better to compensate for market
inequalities. The us, by contrast, lacks an integrated, comprehensive
social-protection system, and has the fourth highest income inequality
of the 28 countries studied. According to us census data, almost 50 mil-
lion Americans could be classified as poor in 2012—a poverty rate of 20
per cent, which is almost double the level in Nordic countries.45

The idea, then, that conditional cash transfers might facilitate a broader
process of redistribution, reducing inequality and all but eliminat-
ing poverty, does not hold in principle, and still less in practice for a
region such as Latin America. Precisely the feature that has made ccts
so popular—their residual nature and cheapness—helps to make them
ineffective in reducing poverty in the long term. Aside from their effects
on income inequality, their emphasis on market inclusion makes it
unlikely they will redress what Göran Therborn has called ‘resource
inequality’.46 Indeed, their very focus on extending commodification
makes them much more likely to compound the vulnerabilities of the
poor, even as state social spending becomes more unbalanced, leaving
them further exposed.

The spread of ccts has now produced a second wave of programmes,

this time financed by private companies and ngos in Africa, which
dispense with the conditionalities, handing over cash without strings
attached (hence the label ucts, Unconditional Cash Transfers). These
are cheaper to run, since they cut the administrative costs of monitor-
ing the programmes and thus reduce bureaucracy. But they also rapidly
incorporate into markets a large mass of people who would otherwise
be unlikely, in the short term, to have access to a stable occupation and
income. ucts have drawn praise from many quarters—the Economist
noted that they ‘work better than almost anyone would have expected’,
and ‘dent the stereotype of poor people as inherently feckless and

45
oecd, Divided We Stand: Why Inequality Keeps Rising: An Overview of Growing
Income Inequalities in oecd Countries, Washington, dc 2011.
46
Göran Therborn, ‘Inequalities in Latin America: From the Enlightenment to the
21st Century’, desiguALdades Working Paper no. 1, 2011.
 40 nlr 84

ignorant’. Nonetheless, ccts have so far been deemed preferable, for

practical and political reasons: after all, ‘people left to themselves may
not spend enough on education or health’, and what is more, ‘attaching
strings reassures middle-class taxpayers that the poor are not getting
something for nothing’.47 The arguments in favour of conditionalities
thus rest not only on their supposed efficacy but also on a logic of control
over vulnerable groups.

The social-protection paradigm that emerged at the end of the 19th cen-
tury and developed, in parallel with the workers’ movements, during the
20th, aimed to protect and equalize access and opportunities, irrespec-
tive of income level and social status. In this model, the structure of
social spending prized not only income security but above all the promo-
tion of equity and convergence. By contrast, the hegemonic paradigm of
the 21st century holds that market mechanisms are the key to improving
general welfare; cash transfers and expanded household debt, the latter
underwritten by the former, are the key elements in this framework,
in which decommodified provision is to be pared to the barest bones.
What is taking place—spurred on by the ‘success story’ of ccts—is a
downsizing of social protection in the name of the poor. Over the past six
years these programmes have benefited from boom conditions, as sur-
plus capital flooded into ‘emerging economies’ from the crisis-stricken
advanced capitalist zones. Yet how they will weather a reversal of capital
flows and tightening of credit, if quantitative easing in the North finally
starts to slow, remains to be seen.

47
‘Cash to the Poor: Pennies from Heaven’, Economist, 26 October 2013.